State-dependent capacitance property and synaptic behavior of memcapacitor based on WS2 nanosheets

Abstract

The limitations of traditional CMOS technology and the von Neumann architecture have driven the exploration of neuromorphic systems, which emulate biological synapses for energy-efficient and fault-tolerant computing. With their simple structure, non-volatile resistive switching, and tunable synaptic weight modulation, memristors are promising components for such systems. Among candidate materials, transition-metal dichalcogenides (TMDs), particularly tungsten disulfide (WS₂), stand out due to their high carrier mobility, strong light-matter interactions, and stability. Recent studies on WS₂-based devices have demonstrated enhanced neuromorphic functionality, though primarily in three-terminal configurations and composite materials. This work investigates a planar Cr/WS₂/Cr memristive device featuring a WS₂ thin layer prepared by a top-down method. Material characterization revealed a single orientation along the z-axis, contributing to excellent self-rectifying analog switching over 1000 cycles and retention time over 10⁴ s. The device exhibits non-volatile and accumulative properties, enabling synaptic weight modulation under ±3V sweeping and 1 ms–100 ms pulse width signals. Additionally, a triangular waveform revealed nonlinear capacitor behavior under saturation conditions, consistent with memcapacitor functionality. Furthermore, the identification of sulfur vacancies as active sites and their reorganization under the applied field clarifies the microscopic switching mechanism, directly linking defect dynamics to the observed memcapacitive behavior. This work provides fundamental insights into state-dependent capacitance and paves the way for designing and implementing next-generation memristive synaptic devices using 2D materials.